The dashboard reports a healthy overall equipment effectiveness (OEE) at the bottleneck. Yet customer orders still miss their confirmed dates. There is no contradiction: OEE describes losses at one resource within a defined time base. By itself, it says little about whether the right products leave the plant on time and in good quality.

Managing a bottleneck therefore requires more than a percentage. Which minute was lost? Which order was affected? Which action changes throughput for the whole process? The preceding insight on production flow explains why high utilization and late deliveries can coexist. Here the focus is daily control of the scarce resource.

Identify the constraint before measuring it

A bottleneck limits attainable throughput for a given demand and period. It may be a machine, inspection station, qualified employee, approval step or material. The constraint may move as product mix changes.

A large queue alone does not prove that a machine is the constraint; early release can create it. Conversely, the most critical machine may stand idle while awaiting the right material. At minimum, compare effective capacity by product family, backlog, WIP history and operational states: producing, setting up, down, waiting for material and blocked by the next process.

A decision-oriented bottleneck KPI set

These measures form a hierarchy. Lost minutes and causes explain an operational change; good output shows the immediate effect; lead time and on-time delivery test whether it reaches the whole system. NIST research describes relationships between manufacturing KPIs. ISO 22400-2 defines formal KPIs for manufacturing operations management; each plant still needs a clear operational time base and data owner.

Why OEE matters but is insufficient

An OEE of 80% can mean very different things. Were planned maintenance and breaks excluded from the denominator? Are microstops recorded? Are defects only detected downstream? Is a shift with many product changes being compared with a long production run?

Rising OEE can even mislead if the machine produces the wrong mix for current customer demand. A sensible changeover can temporarily lower OEE and improve delivery. The daily review should therefore connect OEE components to schedule adherence, good output and orders at risk.

What might one lost minute be worth?

Take a deliberately simplified example. The current bottleneck loses 30 avoidable minutes. At the planned mix it could produce 80 good units an hour. That is 40 units of potential output. At an incremental contribution of €70 per saleable unit, the theoretical contribution at stake is at most €2,800 and requires validation.

This is neither a booked loss nor automatically released cash. The number holds only if demand exists, the units would have been finished and sold without quality problems, downstream capacity is available, and production cannot be recovered later without additional cost. Even so, the calculation helps prioritize the right 30 minutes over hours at a non-critical resource.

Three common management errors

1. Push utilization to 100%. Without an appropriate buffer, material shortages starve the constraint; an unlimited buffer raises waiting time and capital tied up. Set and review a target range based on disturbances and replenishment.
2. Optimize every machine equally. More output upstream often grows the queue. Improving a non-critical resource matters when it protects the constraint, safeguards quality or prevents a new capacity limit.
3. Measure only downtime. An unsuitable mix, frequent reprioritization, rework and late approvals can reduce customer-relevant throughput just as much. The root cause may sit outside the machine.

A practical management cadence

For the first ten working days, record losses at the suspected constraint by shift and reconcile them with orders, buffer and actual good output. Assign owners for recording machine states and resolving differences between ERP, machine and shift data.

Ask five questions in the daily shopfloor review: Which customer order is at risk? Does the constraint have the right material for the next shift? Which loss cause dominates? Who decides on sequence or capacity, and by when? Did yesterday’s action improve good output and delivery?

Review mix, target buffer and constraint location weekly. If demand structurally exceeds effective capacity, volume, delivery-date and investment decisions belong in S&OP/IBP. A shift team cannot close a structural capacity gap on its own.

A useful bottleneck KPI triggers a decision. A useful KPI system then shows whether that decision improved delivery to the customer. The free Production & Operations Health Check helps frame the operational questions; an email account is required only for saved comparisons. Its individual Executive Management Report is €129 net plus VAT. The English S&OP / IBP, Inventory and SCM checks address related decisions.